09 Oct 2023  |   06:26am IST

NIO scientists detect presence of sound making fish species by using AI

AI approach will help in identifying fish species through its sound by interpreting the collected data in real-time; worldwide, 989 fish species from 133 families and 33 orders can produce active sound

SHASHWAT GUPTA RAY

PANJIM: A team of scientists from the CSIR-National Institute of Oceanography (NIO), has detected presence of soniferous (grunting) fish species off Goa coast by studying and matching their sounds using Artificial Intelligence (AI)-based algorithms, which also signify the good health of Goa’s marine ecology.

The four species are Terapon theraps (Tiger Perch, also known as Korkoro), Sciaenidae (Croaker), and Snapping Shrimp. The discovery has been published in the Journal of the Acoustical Society of America. The study was conducted underwater from Grande Island, within the Zuari estuary during the pre-monsoon period.

The findings of the study, which otherwise would have required months to decipher, could be done much faster, using by plugging recorded sounds of the fish species into an AI-based algorithm that correctly identified four species in a matter of minutes.

“Our main objective for AI/Machine Learning (MI) based study was to identify fish sound from recorded data which we had collected using passive acoustic technique. The presence of these fish species in Goa was confirmed using passive acoustic (sensors) technique, combined with AI. We could study the primary sources of biological sounds and differentiate between marine mammals, soniferous fish (sound producing), and invertebrates,” said Bishwajit Chakraborty, former Chief Scientist of NIO, who had initiated this study.

“Using passive acoustic technique, we have extensively investigated this location to understand the underwater environment off-Goa Grande Island area. In the past, Humpback whale sounds were also recorded by CSIR-NIO from this location. Very recently, NIO also recorded fish sound data from the Malvan area off the Maharashtra coast. Using unsupervised machine learning techniques helped us find the success of the number of fish sound identification,” Dr Chakraborty said.

While the main objective of the study was to find out the presence of these fish species, it also helped the researchers to know the health of Goa’s marine ecology.

“The presence of these fish species off Goa coast indicates the health of our coastal reef system and the overall marine ecology. If the coral reefs are dead, then the fish won’t be active and produce sound,” the veteran ocean scientist said.

He further said that more passive acoustic data must be acquired from shallow water regions, especially coral reef regions, to monitor the coral reef.

“Such studies are new in India. Only CSIR-NIO, Goa, and National Institute of Ocean Technology (NIOT), Chennai, is active in this research using passive acoustics. Programmes and funding are needed from various sources for monitoring ocean health, using remote sensing techniques,” he said.